Autonomous Driving Features Promote Driver Distraction
Article: IIHS Study Highlights Need for Robust Safeguards in Partial Automation Systems
The Insurance Institute for Highway Safety (IIHS) has released a study that sheds light on the potential dangers of partial automation systems, such as Tesla's Autopilot and Volvo's Pilot Assist. The research, led by IIHS Senior Research Scientist Ian Reagan, analysed the driving behaviour of volunteers over a four-week period while using these systems.
The study found that drivers are more likely to multitask when using partial automation systems. As drivers grow more comfortable with the technology, the tendency to multitask increases for some, while others are more distracted from the start. IIHS President David Harkey stated that drivers learn to adapt their behaviour to engage in distracting activities when using these systems.
To prevent misuse and ensure attentive driving, robust safeguards have been implemented in these systems. These safeguards include driver monitoring systems, clear human-machine interface alerts, fail-safe and redundancy mechanisms, restrictions on automated maneuvers, limiting system capabilities, regulatory compliance and standard adherence, continuous software updates and recalls, and more.
Driver monitoring systems require continuous or periodic confirmation of driver engagement, such as detecting hands on the steering wheel or eye-tracking to ensure the driver is alert and ready to intervene at any moment. Clear human-machine interface alerts notify drivers when they must take control, especially during system limitations or complex driving scenarios like construction zones, extreme weather, or unexpected traffic hazards.
Fail-safe and redundancy mechanisms in vehicle sensors and power systems aim to avoid single points of failure, including redundant power supplies and interference-free sensor suites (lidar, radar, cameras) to maintain system reliability. Restrictions on automated maneuvers such as semi-automatic lane changes require explicit driver confirmation before execution and limit automation to well-defined conditions where the system is verified to perform safely.
Limiting system capabilities makes it clear that partial automation cannot handle all driving tasks; the system should detect and alert for stationary objects, complex intersections, and other hazards—areas where partial automation is prone to failure. Regulatory compliance and standard adherence, such as following ISO 26262 and VDA 450 safety standards, ensure design practices minimize risk and enhance driver safety.
Continuous software updates and recalls address detected issues like false automatic emergency braking activations or software crashes to reduce dangerous or unexpected system behaviour. These safeguards collectively aim to maintain driver attention and readiness, reduce misuse or overtrust in automation, and improve overall safety during partial automation operation, which still fundamentally requires active human supervision.
The study led by Reagan analysed the driving behaviour of 29 volunteers over a four-week period while using a 2017 Volvo S90 sedan equipped with Pilot Assist. In 72 instances, drivers didn't respond fast enough to prevent alerts from escalating, leading to 16 instances of drivers being locked out of the system. Drivers engaged in nondriving secondary activities, looked away from the road, and had both hands off the wheel more often during the alerts and in the 10 seconds before and after them.
Regardless of the version of the system used, all three groups of drivers engaged in distracting visual-manual activities more often when using Pilot Assist than while driving without it. The percentage of time that drivers were disengaged in the period surrounding the alerts increased, even though the duration of each alert was shorter. Drivers using Volvo's Pilot Assist partial automation system were more likely to check their phones, eat, or do other visual-manual activities while using the system compared to driving unassisted.
The second study, conducted by IIHS and MIT's AgeLab, focused on the driving behaviour of 14 people using Tesla's Autopilot system for a month. Researchers found that drivers using Tesla's Autopilot system quickly mastered the timing interval of its attention reminder feature to prevent warnings from escalating to more serious interventions.
IIHS Senior Research Scientist Alexandra Mueller, the study's lead author, stated that escalating, multimodal attention reminders are very effective in getting drivers to change their behaviour. However, better safeguards are needed to ensure that the behaviour change actually translates to more attentive driving. The studies conducted by IIHS and MIT's AgeLab demonstrate the need for more robust safeguards to prevent misuse in partial automation systems.
- Despite the advancements in technology and automotive industries in the implementation of safeguards in partial automation systems, the finance sector should invest more in research and development to enhance these safeguards, ensuring they are effective in preventing distraction and misuse.
- As the transportation industry continues to rely on partial automation systems, it is crucial for all stakeholders, including finance, technology, automotive, and regulatory bodies, to collaborate and establish standardized safety measures to mitigate the risks associated with driver distraction and the escalating use of automation.
